Which known campaigns qualify?
The view needs a stable inclusion rule and a count of excluded, unknown, or retracted records.
Reward participation · bounded definition
Study campaign cadence
without studying people.
Here, participation means publisher-side reward availability and eligible viewing context. It does not mean that StreamStats knows who watched, earned, claimed, or owns a reward.
Campaign cadence frame
values intentionally absentCampaigns by game or publisher over time
Values omitted
measurement withheld
Earlier periodaggregate pending
Middle periodaggregate pending
Later periodaggregate pending
Claim-window patterns
Deadlines are facts, not behavior.
A future distribution could describe the documented gap between an earning deadline and a claim deadline. It would not say that viewers claimed late, missed a reward, or completed a requirement unless a separately authorized source actually supports that statement.
- Could count
- Source-backed campaigns grouped by game, publisher, or stated period
- Could compare
- Known earning and claim-window shapes with explicit missingness
- Cannot infer
- Personal progress, completion, claim, ownership, or intent
- Cannot merge
- Unknown publishers or games into a convenient guessed identity
Publication questions
must be answered before measurementHow are games and publishers joined?
Only stable, reviewed identities may be grouped; label similarity is not enough.
Which timestamp defines a period?
Earning start, earning end, claim end, and first observation answer different questions.
May the aggregate be retained and displayed?
Collection permission does not automatically grant public display or indefinite retention.
Production boundary
Aggregates wait for a governed data posture.
Real measurement requires GOV-003 to define permitted Twitch Helix aggregate collection, source attribution, observation windows, retention, corrections, freshness, and safe public thresholds. Until that posture is approved, StreamStats publishes no measured result.